Aprizi is about to enter that crucial phase of open beta, so I wanted to pause and write down the latest installment in our customer development journey (part 1 and 2). For anyone new to the blog, Aprizi is building a personalization engine for online shopping — a Pandora for e-commerce, if you will. My thesis is that the intersection of two trends, information sharing and information overload, will enable both the creation and the demand for personalization and intelligent filtering services.
Our customer development journey so far has consisted of the following steps over the last few months:
- 1-on-1 customer conversations with 2 parts: open-ended questions about shopping behavior, and feedback on tangible product ideas (note: I tried one focus group and disliked format)
- an early survey of about 50 consumers
- discussions over paper mockups (really helps focus the conversation when it is time for feedback)
- follow-up discussions around user experience on our crude alpha website
For our “alpha”, we wanted to discover differences between people’s statements and actions. Rather than wait for a lot of technical infrastructure, I asked our alpha testers a few questions and played “personal shopper”, i.e. I manually researched 8 recommendations for each person. Testers then browsed the recommendations on a simple website. Our goal was to test out:
- did people give a damn about this kind of service?
- what kinds of things did they click on and get excited about?
We got great feedback and learned some things that did not come up during paper testing that changed our thinking and approach. For example, we had separated deals vs product suggestions, and had even contemplated a focus solely on deals. Post-alpha, not only did we decide to do both, but we decided not to treat them differently at all.
This alpha process was, of course, entirely un-scalable, but one of the best things we did was start testing with manual labor behind the scenes rather than waiting for automation (Max Ventilla, a co-founder of Aardvark, recently wrote a great post about that as well). After we learned some core take-aways, we ceased alpha testing. My opinion was that we learned what we could even with just a small number of people, and the next real wave of learning would come with beta (which still has manual processes, but will be considerably more scalable).
It is worth noting that I have limited my customer-facing time and focused almost entirely on product for the last few weeks (I do everything but code). If our focus is on learning and validation, and the next big dose of that needs to have the beta in place, then everything else gets second shrift. On that note, every few days Liz and I have been reviewing our coding task list with a scalpel to make sure we’re really stripping out everything but the true necessities. We *have* put time into cleaning up the UI, but we felt that we needed a decent baseline of quality given that our target audience is not techie.
Our forthcoming beta is going to be completely open. I had originally been thinking about doing a gmail-like beta program, where people could invite others, and then I changed my mind. None of my initial reasons held up after a sharp poke with a stick. Here is my thinking:
- we’re not celebrity enough that “exclusivity” will be a huge factor
- we’re not celebrity enough (nor pursuing any PR at this point) that we really need to worry about scalability breakage
- the first version of the app will be crude and bad, but so few people are going to see it, it doesn’t matter from a brand perspective
- far more important is validation and learning, which means maximizing conversion rates, which means letting anyone in who we are blessed enough to *get* to our home page!
Does Lean Startup Work?
What have we gained over the last 4 months in getting to this point? It is too soon to say what our “product market fit” will look like, but I can say than thinking about lean startup ideas has sped us up. It helped us focus on the most important things, which is *beyond* critical when you are a tiny, bootstrapping team. When your brain is swirling night and day around customers, market, distribution, partners, monetization, competition, investment, etc. it is great to have the lean-startup framework to come back to and cut through the mental noise.
(btw, I think the “lean startup” framework should be judged on how it enhances speed, prioritization and focus, not the ultimate success or failure of a particular company)
Have we done everything perfectly? No. We could have done some things faster, and we lost time early in the process validating some technical hurdles rather than ruthlessly focusing our coding effort on customer value proposition. I also think I should have prioritized landing page tests earlier. But we’re here, we’ve learned a lot, and we’re about to learn a ton more, good or bad.
If you want to get an email when open beta kicks off (and I hope you do!), please sign up over at Aprizi or just write me directly (giff.constable -at- gmail).